[mmaimcal] Receiver PDR notes

[Al Wootten]

Jeff, Simon, Peter, Mark were there for part of the meeting.  Here are
some extracts from my notes which I thought ImCal might have an interest in.
Opinions are my own!  Corrections welcome, as reconstructing notes now for
digestion from a meeting a month ago is not easy.

Front End PDR Meeting in Tucson.  Some notes:  
Optics:  Peter noted that
the specs on optics tilt didn't consider the possible effects of the 
introduction of a phase gradient across the primary beam; antenna to antenna
variation could prove a nasty problem.
Lamb noted that cancellation of cross-polarization does not work with a grid
between mirrors, as with one band 7 design.  reduction of the angle of
incidence on that mirror will address this some.  It may be better to
concentrate polarization work at the high end of band 6 (which does not have
a grid) rather than in band 7 if this desgn were adopted  Band 3 beam squint
is estimated at 7%, same as the VLA Peter says.

Calibration:
Plambeck noted that the saturation T for the ALMA receivers will probably be
about 2500 K at 110 GHz as opposed to 300K for BIMA.  Nonetheless how do
we calibrate?
a) Decree that the gain is linear.
b) Use Cold loads.  But the tradeoffs here have been discussed and this option
     discarded.
c) Chopped noise source a la BIMA.  This adds noise to the system temperatures
we have valiantly minimized.  The coupling is hard to measure accurately, as
has been noted.  Furthermore, truncation by optical elements makes coupling
frequency dependent.  Scattering at edge of the beam adds in or out of phase
with respect to the center according to frequency.
d) Semi-transparent vane.  The absorption is calibrated on the telescope.
Treatment so far assumes the loss is resistive.  Could phase errors across
the beam be introduced?  

There should be a central weatherstation to measure Tm with e.g. a 50 GHz
radiometer, tau with a tipper, tau(freq) with an FTS, and a 183 GHz WVR to
track tau at each antenna.

We should prototype high quality ambient loads and semi-transparent vanes
in the widget area.  

We should test the subreflector noise source concept on the prototype with
careful attention to variation of performance with frequency.

WVR (Hills):  There will be a WVR PDR sometime in the summer 2001.  The project
started 15Jan01, one each Dickey and correlation rx will be built.  
Apropos Real time refreaction correction:  To get
a gradient across the antenna we might need 4x the sensitivity, marginal
with the Dickey-switched rx.  Hills showed a design for pointing system
optics which fits into the space between cryostat and top of receiver cabin.
Objections were noted at hogging all the widget space.  Baseline WVR System 
ready Jan03.  10mW of 90GHz doubled to 183 GHz, tripled power rejected with rf
bandpass filter.

Receivers:
Band 3 and 6 look like they are coming along well to me.  A decision needs
to be made on HFET or SIS for Band 3 soon; SIS development continues.  In
general, there was a lot of whining about the difficulty of achieving the
stability requirement but I heard very little in the way of plans to try
to achieve it.  There was some discussion of sideband separation.  I pointed 
out that ALMA will work with TP and correlated data which makes SB separation
via e.g. Walsh functions an incomplete solution.  SB separation
might be measurable interferometrically but unless single dish observations
are done very synchronously with the interferometer observations I don't see
how we can accurately calibrate the SD data easily.  Perhaps somehow via
injected signal.  But SSB operation is the best solution, IMHO, in terms of
noise and evaluation of the data.

Belitsky (OSO) presented a band 7 design 2SB at 4-8 GHz IF (I think).  Just 
beginning stages of design but IMHO he is on the right track.

Claude (IRAM) presented a fallback design - DSB mixers 4-8 GHz IF and discussed
a development task to produce balanced SB separating mixers 4-8 GHz IF (2SB).
Grids in the present design may compromise polarization work. TSSB 23-35 K
expected at 275 - 370 GHz.

Problem:  Both Band 7 designs are for 4-8 GHz IF while the LO is designed
for 4-12 GHz.  This is the widest band; some loss of frequency real estate
may be necessary.  Where?  IMHO, lose it at the top end first, though I
dearly love them ol' H2CO lines.

Baryshev (SRON) presented a single ended Band 9 DSB receiver could be 4-12 GHz 
IF or 4-8 GHz IF.  Balanced mixer tests Apr01.  Higher IF bandwidth will 
only be pursued if it can be
achieved without undue increase in receiver noise temperature and instability
with respect to the narrower IF bandwidth.  Basically, the JCMT D band rx is 
baseline design I think. TDSB~350 K.  Development is a single ended 
quasi-optical mixer but SB separation is not in the cards.

Sekimoto:  Band 8 DSB dual polzn for evaluation Apr2003 on ASTE.  ASTE can
use ALMA inserts, so could be used for high frequency evaluation of receivers
pending arrival of production atennas.  They need an ICD very soon.

A problem was noted in that the multiplier from CDL for band 9 isn't expected
until late 2002.  Possible to use limited range Zimmerman multiplier until
final ALMA multiplier is available with restricted frequency range.

I'll post the official committee recs at next week's ImCal meeting agenda.

Al